The effect of cue-criterion correlations, cue intercorrelations and the sign of the cue intercorrelation on performance in suppressor variable tasks.
Kerstin Armelius, Bengt‐Åke Armelius · PsycEXTRA Dataset · 1975
Arraelius, K., and Armelius, B. The effect of cue-criterion correlations, cue intercorrelations and the sign of the cue intercorrelation on performance in suppressor variable tasks.Umeå Psychological Reports No. 81, 1975.-The subjects' performance was in vestigated in a two-cue MCPL-experiment with a 2 (Cuecriterion correlations, r^ = .60 and .80) a 2 (Total task predictability, R* = 1.00 and .70)x 2 (Sign of r..) x 5 (Blocks) factorial design.The difference in /J 2 in R was due entirely to an increase in the cue 2 intercorrelation, r^.Effects of Rg on performance were therefore interpreted as effects of the cue intercorrelation.Two control conditions, with orthogonal cues and the same values of rg^ as in the experimental conditions, were included in the design.Subject consistency was directly related to the cuecriterion correlations and the cue-judgment betaweights were directly related to the magnitude of the cue intercorrelation.The sign of r.. determined the 13 level of consistency.In multiple-cue probability learning (MCPL) tasks the subjects are required to make predictions about a criterion variable from the state of a number of cue variables.When cues are orthogonal the only source of information about the validity of a cue is the cue-criterion corre lations, re£.It is well known that the subjects are able to learn to utilize the cue-criterion correlations in orthogonal MCPL-tasks, when they make their predictions (see e.g., Dudycha & Naylor, 1966).